Conventionally, transducer designers have used epoxy resin adhesives to join stacks of ferroelectric rings to head and tail masses with the goal of achieving a low-loss coupling between the active element and the radiating surfaces. Better operation characteristics and avoidance of the generation of self-destructive tensile forces have dictated that longitudinal compressional prestressing be included to help the coupling among elements. In addition, it is known that a thin, rigid glue line optimizes the impedance matching across the joints between the active and the passive transducer elements. However, the different thermal expansion coefficients of the dissimilar materials often failed in fluctuating temperatures and demonstrated the unsuitability of the rigid thin-line joints. To elaborate, a rigid thin glue joint between an aluminum head mass and a piezoelectric ceramic stack often tore the stack apart during curing. Since the thermal expansion coefficients of these materials are 23.8.times.10.sup.-6 cm/.degree.C. and 3.8.times.10.sup.-6 cm/.degree.C. respectively, the elevated curing temperature of epoxy resin adhesives ranging from 150.degree. to 200.degree. F. creates an intolerable stress level. Similarly, self-destructive stress levels are set up when the transducer is operated in a cold ocean environment, for instance, under an ice pack. Unfortunately, the rigid epoxy resin adhesive does not bend or give as the stack and head elements undergo their different ranges of flexure but rather the ferroelectric or ceramic element is torn apart due to its inherent low tensile strength. Furthermore, the brittle ceramic element shatters if the transducer is subjected to shock because the rigid epoxy resin joint transfers all the impact to the fragile element. Another disadvantage of using an epoxy adhesive joint is its permanent nature. For example, when different operating characteristics are desired, it is expedient to change the head and tail mass and the costly ferroelectric stack unavoidably is destroyed as the masses are being removed. One notable attempt at remedying the enumerated shortcomings of a rigid joint employs a rigid fiberous glass-epoxy shim to couple the ferroelectric driving element to the metal head and tail masses. Using stress rods to hold the parts together does provide a decoupling across the joints and does solve thermal expansion problems. The main deficiency of this approach becomes apparent when such a transducer is operated for the joints do not mate intimately and the impedance match, necessary for responsive operation, is lacking.